Negative Feedback Gain Control

[jnb]

Originally posted by AndrewT
C605//C607 are the DC block on the NFB return.
They increase the DC impedance to near infinity with the result that the DC gain falls to near 1times (+0db).

Could these be made smaller, to limit the low frequency response of the amp (like a filter) and to get away with a smaller higher quality cap? What issues could this create?

[AndrewT]

Originally posted by jnb
Could these be made smaller, to limit the low frequency response of the amp (like a filter) and to get away with a smaller higher quality cap? What issues could this create?

you can, but you shouldn't.
Use the input filters to control the bandwidth of the amplifier.

[jnb]

Thanks, Andrew, I will.

[kingneb]

Hello,

I am setting up a system for a master gain control for a preamp. Digital potentiometers will be used to set the gain. There will be a traditional log volume pot (digital) on the input. In the feedback loop there will be another digital pot that varies to set the gain from about unity up until 20db max. Should I use a linear or a log pot in the feedback loop?

[unclejed613]

log. it works the same whether it's in the feedback loop or signal path.

[AndrewT]

?
Are you sure?
I have not done the sums, but I suspect that a linear pot in the NFB loop will not give a linear change in gain. I suspect it will behave/control more like a audio taper vol pot.

[unclejed613]

there are preamps that use the same method for the main volume control. they are audio taper pots.

[AndrewT]

You are forcing me to prove the point.
I am in the middle of doing the sums.

[kingneb]

In a simulator the gain is non-linear with a linear pot. It resembles an exponential increase. At first there is barley any change in gain but takes off in the end. With a log pot the effect is exacerbated even more.

The picture shows an example of this situation using a linear pot. What I am trying to do is create a variable gain control, in an opamp feedback loop, to adjust the master gain of a line stage.

The intent of this exercise is to compensate for signal level disparities in various audio components that supposedly put out a "line level" signal. One line component may put out a lower or higher signal relative to the nominal line level rating of .447 volts peak. Also, different power amplifiers want different signal levels. So to ensure an even balance between all stereo components, this master gain is adjusted and stored in memory for each selected component.

Based on the chart, is a linear pot the way to go here?

[AndrewT]

Originally Posted by kingneb

In a simulator the gain is non-linear with a linear pot. It resembles an exponential increase.

That makes sense.
Looks like I don't need to prove it to you.